Affordable Access

Access to the full text

Decoding of Pain Perception using EEG Signals for a Real-Time Reflex System in Prostheses: A Case Study

Authors
  • Tayeb, Zied1
  • Bose, Rohit2, 3
  • Dragomir, Andrei2, 4
  • Osborn, Luke E.5, 6
  • Thakor, Nitish V.5, 7
  • Cheng, Gordon1
  • 1 Technical University of Munich, Arcisstraße 21, München, 80333, Germany , München (Germany)
  • 2 National University of Singapore, 28 Medical Dr. 05-COR, Singapore, 117456, Singapore , Singapore (Singapore)
  • 3 University of Pittsburgh, 3700 O’Hara Street, Pittsburgh, PA, 15261, USA , Pittsburgh (United States)
  • 4 University of Houston, 3517 Cullen Blvd, Houston, TX, 77204, USA , Houston (United States)
  • 5 Johns Hopkins School of Medicine, 720 Rutland Ave, Baltimore, MD, 21205, USA , Baltimore (United States)
  • 6 Johns Hopkins University Applied Physics Laboratory, 11100 Johns Hopkins Rd, Laurel, MD, 20723, USA , Laurel (United States)
  • 7 National University of Singapore, Engineering Drive 3, 04-08, Singapore, 117583, Singapore , Singapore (Singapore)
Type
Published Article
Journal
Scientific Reports
Publisher
Springer Nature
Publication Date
Mar 27, 2020
Volume
10
Issue
1
Identifiers
DOI: 10.1038/s41598-020-62525-7
Source
Springer Nature
License
Green

Abstract

In recent times, we have witnessed a push towards restoring sensory perception to upper-limb amputees, which includes the whole spectrum from gentle touch to noxious stimuli. These are essential components for body protection as well as for restoring the sense of embodiment. Notwithstanding the considerable advances that have been made in designing suitable sensors and restoring tactile perceptions, pain perception dynamics and its decoding using effective bio-markers, are still not fully understood. Here, using electroencephalography (EEG) recordings, we identified and validated a spatio-temporal signature of brain activity during innocuous, moderately more intense, and noxious stimulation of an amputee’s phantom limb using transcutaneous nerve stimulation (TENS). Based on the spatio-temporal EEG features, we developed a system for detecting pain perception and reaction in the brain, which successfully classified three different stimulation conditions with a test accuracy of 94.66%, and we investigated the cortical activity in response to sensory stimuli in these conditions. Our findings suggest that the noxious stimulation activates the pre-motor cortex with the highest activation shown in the central cortex (Cz electrode) between 450 ms and 750 ms post-stimulation, whereas the highest activation for the moderately intense stimulation was found in the parietal lobe (P2, P4, and P6 electrodes). Further, we localized the cortical sources and observed early strong activation of the anterior cingulate cortex (ACC) corresponding to the noxious stimulus condition. Moreover, activation of the posterior cingulate cortex (PCC) was observed during the noxious sensation. Overall, although this is a single case study, this work presents a novel approach and a first attempt to analyze and classify neural activity when restoring sensory perception to amputees, which could chart a route ahead for designing a real-time pain reaction system in upper-limb prostheses.

Report this publication

Statistics

Seen <100 times